Marketers adopted AI analytics and automation faster than almost any tool that came before it, and now fewer of them can prove it worked. A survey of 1,400 marketing professionals run by Jasper for its State of AI in Marketing 2026 report found AI use jumped from 63 percent to 91 percent in a single year. Over the same period, the share of marketers who said they could demonstrate AI return on investment fell from 49 percent to 41 percent. Adoption nearly tripled the prior pace of growth. Proof of value went the other direction.
Part of the gap traces to what gets measured in the first place. Jasper found the top metric marketing teams track for AI is hours saved, not revenue, pipeline, or any outcome a finance team would sign off on as return. Ninety five percent of the marketers surveyed said they plan to increase AI spending in 2026 regardless, which means budget is scaling well ahead of any agreed way to check whether the spending is working.
The data these systems run on has not caught up either. Validity surveyed 500 B2B and B2C marketers across the US, UK, Brazil, Australia and New Zealand for its State of CRM Data Management in 2026 report and found only 21 percent consider their CRM data very well prepared for AI use. Despite that, two thirds of the same organizations increased the number of marketing decisions delegated to autonomous AI agents over the past year. Validity also found 78 percent of C-suite respondents and 92 percent of SVP and VP level respondents had already acted on an AI recommendation they later suspected was wrong because of bad underlying data, and 62 percent of organizations said they lost revenue directly from poor CRM data quality.
Gartner points at governance, not the AI itself, as the reason this keeps happening. Shiva Varma, a senior director analyst at Gartner, said organizations tend to treat AI agent governance as binary, either locked down or fully trusted, which he called the root cause of failure, according to Gartner’s analysis reported by The Register. Gartner predicts 40 percent of organizations will demote or decommission AI agents by 2027 for that reason: agents given too little oversight create compliance and security risk, and agents given too much slow everything down and push teams toward shadow workarounds instead.
Put the three findings together and the declining ROI number stops looking like a mystery. Marketing teams are handing more decisions to AI systems built on CRM data that four in five of their own executives do not fully trust, while judging success by time saved rather than outcomes a finance team would recognize. That combination makes a falling proof of ROI rate the expected result, not a surprise. The organizations due to walk back their AI agents by 2027, on Gartner’s timeline, are likely the ones that scaled the automation before they fixed what it was reading from.
Sources referenced in reporting: Jasper, The State of AI in Marketing 2026, Validity, The State of CRM Data Management in 2026, The Register, 4 in 10 AI agents headed for demotion or the rubbish bin.